the Value of Kd suggests it is good enough for a stable complex formation. One of our Protein-DNA complex had Kd of 200nM and we separated the complex from SEC. Somebody told me a while back that Tris Buffer is not a good buffer to keep protein as it has primary amine (nucleophile) and it may hinder mask the important phenomenon. I would just use phosphate buffer. In addition, i dont know if SPR gives the stochiometry of the interaction and if not then do ITC and find out the stochiometry and set crystal trays just by mixing appropriate amount of these two proteins.

Is it possible that the complexes under study contain one or both of the components need to be modified (glycosylated, pgosphorylated or acetylated)? The interactions you measured by SPR analysis may be the tethering of the molecules rather than stable complex formation. In addition, protein complexes often contain one or other components which escaped your initial discovery. These points of concern can be easily analyzed by protein-protein interactions in vitro by native PAGE with labelled components, which will enable you the stoichiometry and conditions of complex formation precisely before the chrystallization trials.

If one of your proteins is tagged (His for instance) you might mix both with the untagged one in excess and load it on NiNTA (IMAC). After an extensive wash you could elute and see if the partner coelute with your tagged protein. If yes, load on SEC (adjust ionic strength to at least 0.15 mM NaCl) to see what are the molecular weights of the eluted species.

Estimating binding constants is usually straightforward. Complex reconstitution and purification may be tricky. Before starting crystalisation assays you should insure that you are able to purify complexes.

Lastly I would recommand HEPES or MOPS instead of Tris. Tris pH vary considerably with T° (0.3 per 10°C). Phosphate is good to providing you do not use divalent cations (Ca++ especialy complexes with phosphate).

Before you try and crystallize it is good idea to do size exclusion. Immediately after this test on SDS and Mass-spec for degradation products and impurities. Check how stable is your sample in high concentrations does it precipitate. If possible do 1d-nmr or CD to check if it is structurally defined.

May be go and speak with crystallographers in Lund.

are these proteins monomeric in solution? With such tight Kd, you should see only complex at typical working protein concentrations, however.. What were the designs for SPR experiment for Kd determination?

What do you mean by you were not able to crystallize the complex? Did you get crystals of only one protein? Or did you just not get any crystals?

Thanks for suggestions.

We recently solved structure of our bacterial protein that binds to eukaryotic protein, but do not make complex with eukaryotic partner. The eukaryotic protein is expressed in HEK293T cells and purified. It is glycosylated and we compared its behaviour with commercially available protein. It is same in all properties. we cant detect complex in gel filtration and native PAGE, if we mixed up both proteins. Now we will make some changes in buffer, let see if it works.

Thanks for your valuable inputs.

Hi,

If you do not see any complex after SEC, but see complex formation during SPR, hm, this is something weird. If you immobilize either of them on a chip and analyze binding of another one, you always have similar Kd in SPR. My suggestion would be to do ITC and if they bind with nanomolar affinity, they should do during ITC. After ITC will tell you that you have a complex, you can save this sample and crystallize it.

Did you try to express (transfect) the bacterial protein in HEK293T cells, assuming that this bacterial protein is a soluble effector protein for example. I do not know the nature of the proteins you are studying, so that I would help more and this is what you might have to really consider how do you plan your experiments. Anyway, I wish you good luck in crystallizing the complex.

Andrei

Why not just express the eukaryotic protein in bacteria and the 40nM kd complex should form. We did this with 2 mammalian proteins of similar kd and formed a complex that we eventually crystallized.

Michael Sacher has a very good idea to express both proteins in E. coli where there is no possibility of post translational modification other than removal of MET from N-terminus. Most likely the partner protein being expressed in a eukaryotic cell line is heterogeneously glycosylated and inhibiting the crystallization of the complex because even if the complex is forming it is heterogeneous and will never crystallize.

We performed the host protein expression in E. coli and it is going streight to inclusion bodies, we didnt use this protein upto now. Our bacterial proteins is also going to inclusion bodies. So finally we are purifying them from Urea. Now may be if we mix both inclusion bodies and refold them together, may be they will make complex! We will definitely try it.

Thanks,

Birendra have you tried inducing the expression of the protein at 20 degrees celcius instead of at 37? In my lab we have had alot of sucess using this protocol where the protein is insoluble at 37 degrees (in inclusion bodies) but completely soluble at 20. We grow the cells in LB broth at 37 degrees to an OD600 of 2.0, cool the cells at 20 degrees for 30 minutes, induce with 1 mM IPTG and let them grow overnight at 20 degrees and harvest them in the morning.

Do you have any idea on the temperature at which this proteins are stable? if you don't, i think you need to get their temperature of stability using enzyme kinetics

Dear Mr. Birendra Singh,

Do you have the crystall structure of the isolated proteins? If yes, you can try the following:

1) If the glycosilation is not important for the interaction you can try some constructions in E. coli. You express them with a large and flexible linker (ex: GGGGG... or GSGGSGGSG...) in two constructions: A-link-B and B-link-A. You can also try more then one linker.

2) If you have the two solved crystall structures, you can suppose how they interact and can force some chemical bond (ex: S-S bond by reductor agent). It will increase the time of interaction, since your kD is so weak.

Another suggestion do not depend on knowledge about the crystal structure:

- You said you can isolate the complex by gel filtration. Have you tried the crystallisation with this peak of gel filtration?

Good luck!

Success in your experiments.

Another thought - I agree that knowing the stability of the complex once formed is important to this effort given the timescale for crystallization. Also, are you certain that this protein-protein interaction does not involve the native glycosylation moiety? If it does, then differential glycosyation from the overexpression system will be a problem and may even inhibit crystallization (and you could still get a Biacore signal from association of a subpopulation). Also, are you certain that your glycoprotein is not undergoing oxidation during your crystallization and so disrupting a critical interaction (maybe try adding a bit of DTT)? I also agree that reducing the rate of expression in your bacterial co-expression system (through reduced temperature, less rich culture medium - possibly defined medium, or better control of induction - titration of IPTG or use of a different promoter) is a good strategy.

Protein - Protein interaction affinity depends on ionic strength. IF the interaction surface is hydrophobic you could stabilize the complex with increased ionic strength and vice versa. To know what type of interaction you have, you can do ITC and different salt concentrations and then decide how to set up crystallization of the complex.